Action Capsules: Human Skeleton Action Recognition

TL;DR

This paper introduces Action Capsules, a novel method for skeleton-based human action recognition that effectively captures joint dependencies and outperforms existing approaches with lower computational costs.

Contribution

The paper proposes Action Capsules, an end-to-end network that identifies key joints and encodes their correlations for improved action recognition.

Findings

Outperforms state-of-the-art on N-UCLA dataset

Achieves competitive results on NTURGBD dataset

Has significantly lower computational requirements

Abstract

Due to the compact and rich high-level representations offered, skeleton-based human action recognition has recently become a highly active research topic. Previous studies have demonstrated that investigating joint relationships in spatial and temporal dimensions provides effective information critical to action recognition. However, effectively encoding global dependencies of joints during spatio-temporal feature extraction is still challenging. In this paper, we introduce Action Capsule which identifies action-related key joints by considering the latent correlation of joints in a skeleton sequence. We show that, during inference, our end-to-end network pays attention to a set of joints specific to each action, whose encoded spatio-temporal features are aggregated to recognize the action. Additionally, the use of multiple stages of action capsules enhances the ability of the network to classify similar actions. Consequently, our network outperforms the state-of-the-art approaches on the N-UCLA dataset and obtains competitive results on the NTURGBD dataset. This is while our approach has significantly lower computational requirements based on GFLOPs measurements.